DATA COLLECTION & SAMPLING

Transcription

1 DATA COLLECTION & SAMPLING The Local Landscape Fall 2017

2 Preparing for data collection

3 Increase background knowledge Research Plan Field Recon Test field methods

4 Field reconnaissance Visit geographic location to walk around and observe Pilot study initial run through of data collection methods

5 Research plan Operational definitions Geographic units Defining Study Area Literature Review Base Map Sampling

6 Getting data Where can I find GIS/mapping data? Portland Civic Apps Metro RLIS (Regional Land Information Systems) Oregon Geospatial Data Library US Census Bureau Simply Map

7 How much data is enough? Sampling

8 What is a sample? A subset of the population for your study area/research It is not always (usually) necessary to collect information from every person in the community Subset should accurately represent the characteristics of the total population

9 What is a population? The aggregate or totality of individuals about which inferences or generalizations are made from your research

10 Sampling in qualitative research Samples sizes are usually small Keep collecting data until nothing new comes out of the data

11 What determines sample size? Based on: Research goals and objectives Characteristics of the population (size and diversity)

12 Research objectives & sampling Three distinct types of objectives: to describe the phenomena occurring within a specified area, to develop general principles of location, to apply the principles of geography to a specific problem

13 To describe Does not generally sample because the goal is to enumerate as many phenomena as possible

14 To develop general principles Requires the researcher to know how the specific subset of individuals (objects, persons, events, areas) are related to the total population about which principles will be developed

15 To apply Ensure that the specific study area/sample is a member of the population to which the general principles pertain

16 Types of Sampling

17 Random sampling To keep sampling unbiased, random sampling gives each individual of the population an equal chance of being selected Example: Choosing every 5 th house on the block to interview or survey

18 Unrestricted, random sample n units are obtained from a population of N units in a way that each unit has an equal and independent chance of being chosen How is this implemented? numbering the population units from I to N, and then drawing a set of random numbers from a collection of numbers 1 to N "Drawing" the random numbers may be carried out by picking numbered papers from a bag, using a Table of Random Numbers or by generating a set of random numbers from a computer

19 Haphazard/Accidental sampling Choosing individuals arbitrarily without any conscious attempt by the researcher to pick particular members of the population Factors that affect human decisions invariably create a biased sample

20 Purposive sampling Select participants that are likely to generate useful data for the project Ensure that the you choose an indicative sample that is seen as credible - statistical representativeness is not the aim Minimize sample bias Recognize that the sample will not represent the entire population, therefore you cannot make claims for the whole population

21 Purposive sampling Groups participate based on a preselected criteria relevant to the research question Example: Places that offer services to homeless Sample size depends on: Study area Time Resources Theoretical saturation (the point when new data no longer brings additional insights)

22 Quota sampling When designing the study, decide how many people with which characteristics to include as participants Characteristics might include age, place of residence, gender, profession, marital status, location, size, etc. Example: Wild spaces in Portland 5 from each quadrant of the city, 2 of which are larger than 10 acres. Restaurant preference 10 males, 10 females, half of which are between and the other half 30+

23 Snowball sampling Participants with whom contact has already been made use their social networks to refer the researcher to other people who could potentially participate Snowball sampling is often used to find and recruit groups not easily accessible to researchers through other sampling strategies

24 Overview of sampling strategies Type of Sampling Intensity sampling Deviant case sampling Stratified (Quota) sampling Purpose To provide rich information from a few select cases that manifest the phenomenon intensely but are not extreme cases To learn from highly unusual manifestations of the phenomenon in question To illustrate characteristics of particular subgroups of interest; to facilitate comparisons Example Interviewing LBGTQ homeless to understand how they came to be homeless Interviewing people that only use plastic bags in a community where it is common for people to typically use re-usable ones Collect data about restaurant preference from different sexes, age groups, and socioeconomic status

25 Overview of sampling strategies Type of Sampling Snowball or chain sampling Maximum variation sampling Convenience sampling Purpose To facilitate the identification of hard-to-find cases To document diverse variations; can help to identify common patterns that cut across variations To save time, money and effort. Information collected generally has very low credibility Example Finding out about other graffiti artists in town by asking the few that you already have identified Looking at the use of paper v plastic v reusable bags by going to stores of varying sizes, in varying locations, and selling a variety of goods If going door-to-door interviewing, using a sample based on who answered the door and your questions the day you went out to collect data

26 Overview of sampling strategies Type of Sampling Criterion sampling Purpose To investigate in depth a particular type of case; identify all sources of variation Example Interviewing only people who have filed a noise complaint about the trains